Yep, exactly, but defo not illegal at all
Not possible , it has to be an intraday tradeSure. Just leave them there forever, and one side will always be winning, with the other always losing.
that is the problem . its hard to do the position sizeThis is what I do, mostly. The trick is to cut loses on the losing one and keep the winner running. Easier said than done, but you have something on your side: Trends normally continue steady for a couple of days, so there is that. On average if you keep the winner running for longer than the loser, it should generate a positive net amount. The result is a surviving algo that yields very little profits. Better than nothing.
that is the problem . its hard to do the position size


George Soros would take inverse positions at times - just some food 4 thought.
"As a general principle, I do not dismantle positions that are built on a thesis that remains valid; rather, I take additional positions in the opposite direction on the basis of the new thesis. The result is a delicate balance that needs to be adjusted from moment to moment. If the balancing act fails to protect the portfolio, I have to cut back all around in order to assure survival. If the balancing act is successful, I gain liquidity without having to sacrifice desirable positions. To illustrate: if I start with a fully invested position and then sell short an equal amount, a 20% decline, even if it affects the longs and the shorts equally, leaves me only 80% invested on the long side. If I cover my shorts at the right time, I come out way ahead, but even if I cover my shorts with a loss I am better off than if I had sold my longs at the wrong time. In practice, the operation is much more complex because the balancing act is not confined to the stock market."
George Soros
%%It doesn't matter what the position size is. Whether 1 contract or 1000. All you're doing is what I call a PnL lock. You're capturing a move for a win no matter which direction the thing goes. As someone else mentioned up above, the real trick is when to close the losing leg.
Ultimately all you're doing is trying to time the peak or trough of a cycle, which you might as well do with just one leg so you have half the trading expense.

]